Messages in this thread | | | Date | Mon, 31 Jan 2022 09:23:20 -0800 | From | "Luck, Tony" <> | Subject | Re: [PATCH v2 0/6] PPIN (Protected Processor Inventory Number) updates |
| |
On Mon, Jan 31, 2022 at 01:31:49PM +0100, Borislav Petkov wrote: > On Fri, Jan 21, 2022 at 09:47:37AM -0800, Tony Luck wrote: > > ... > > They look good so far on my PPIN-enabled AMD box. > > > 5) Add "ppin" to /sys/devices/system/cpu/cpu*/topology/ppin > > > > The big question for this part is whether there is a better > > place to expose this value. I'm open to other suggestions. > > Yeah, I'm not sure about that either. I have > > $ grep -r . /sys/devices/system/cpu/cpu*/topology/ppin | cut -d: -f 2 | uniq -c > 32 0xxxxx > > 32 times the same number. > > Wouldn't > > /sys/devices/system/node/ > > be a better place?
Maybe.
> Even if those were logical nodes, it would still be less needless > replication and that would be one more way for root to figure out which > logical nodes belong to the same physical package... :-)
That would work for existing products. There are some cases where a single package appears as mutiple nodes. But that's ok ... as you say, one more clue to the topology.
I'm worried that some future thing might reverse that and have a "package" id for each die in a multi-die package which still appears as a single node. That would distort the meaning of "package", so it isn't supposed to happen. But if it did, Linux would be stuck just reporting one of the "package" ids.
-Tony
| |